You are correct that there is not much recourse when you have a tape
error during an import.
Export/Import via server-to-server is a hard way to go with extremely
large amounts of data, which is why, as others have said, sites
frequently choose to NOT transfer the old backups.
You are NOT going to make this work with a 100MB connection, period.
(Would you expect to FTP a TB of data over a 100MB connection? I think
NOT!)
If your management wants the data transferred, your choices are
1) pay for a faster connection
(And you'll STILL have problems getting it done server-to-server - would
you want to FTP a TB of data over a 1Gb connection? Still NOT!)
2) Buy, beg, rent, borrow a standalone 3590 external drive to connect to
your AIX box, configure as a manual TSM library.
(You'll have a good chance of borrowing one from the IBM'er who sold you
the AIX box.) Export to 3590 cartridge from Z/OS, import from the
standalone drive to AIX. Much faster. Much better solution for bulk
data.
Wanda Prather
"I/O, I/O, It's all about I/O" -(me)
.
-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Zoltan Forray/AC/VCU
Sent: Thursday, January 20, 2005 10:17 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Moving clients between TSM servers
I am looking for suggestions on how to handle moving large TSM clients
between TSM servers and maybe some corrections to my understanding of
how
EXPORT/IMPORT handles error situations.
I need to move TSM clients from my zOS server (being phased out) to my
AIX
TSM server. The problem is, some of these systems have > 1TB of data to
move.
When I tried to move a 500GB client (250GB actual since the copypool
data
is not moved), the export/import ran for >40 hours before failing on a
tape drive failure (190GB moved).
My understandings are that:
1. During the E/I the client is locked/can not be performing backups on
*EITHER* system. This means the client has to go without *ANY* backups
for
the duration. This is generally not acceptable !
2. If the E/I fails like it did, I have no recourse but to delete what
was successfully sent, and then start all over again. Doing a MERGE
FILESPACES is only going to save me the issue of it creating multiple
duplicate directories with the same name +1. It still will work through
all 40+ 3590 tapes.
So, how do you handle situations like this ?
If 190GB took 2-days, 1TB will take > week with me praying for no tape
drive failures (or TSM server crashes like it did earlier in the week
with
a U301 abend/known problem).
The zOS server only has access to a 100MB connection.
Both TSM servers are 5.2.3.4. I will be upgrading to 5.3 as soon as I
can.
|
